3D Printing Waste Reduced with AI Planning
Why this is here: The research uses AI to address a specific manufacturing problem—reducing plastic waste—and provides a mathematically sound basis for improving efficiency in 3D printing.
A new artificial intelligence system cuts down on wasted plastic in 3D printing. Researchers studied how to best assign rolls of filament—the plastic “ink” for 3D printers—to different parts being made on demand. They focused on “make-to-order” systems where printers quickly switch between projects.
The team modeled the process as a complex math problem called a Markov Decision Process. This helped them understand how to make good decisions about which filament roll to use for each part.
They found a simple rule, an “index policy,” that improved on random assignment. This rule scores each roll based on how much waste it would create.
They then combined this rule with “deep reinforcement learning,” a type of AI. This AI learns from experience and fine-tunes the assignment process.
Simulations and real-world tests showed the AI significantly reduced wasted filament. The system is fast enough to work in real time and provides clear reasons for its choices.